I need to allocate 'n' times a structure in memory.
Like 'n' * sizeof MYSTRUCT
Im ok on how to do the allocation, but not on the multiplication to arrive at the correct
allocation size. ie: I cant just do a mov eax, var * sizeof MYSTRUCT
So how do I do this ?
rgs, striker
A couple of questions, how many times must you do the calculation . If its within loop code I would be looking for a fast way to do it and this will depend if its a constant multiplication factor or not. Shifts, LEA and ADD are the type of instructions that come to mind if its a constant, if not, you can use the integer capacity of FP code to do a simple MUL or code a normal integer mul with MUL or IMUL.
I do it once, and currently I do this....
mov eax, sizeof STRUCT
mov ecx, count_of_items_i_want
mul ecx
; go on and allocate 'eax' bytes
Do this appear correct ?
rgs, striker
I doubt that you will improve on that for a variable size mul.
Hi,
If the size of your array is a power of 2 (like 2, 4,8,16,32 etc). You can use shifts instead. SHL for multiply, SHR for divide. For example for multiplying eax by 8 do
shl eax,3 ; 8 = 2^3
and for multiplying by 32
shl eax,5 ; 32 = 2^5
Thomas :U
Hutch, could it be any faster if the math were "offloaded" to the ALU? Maybe this is a question for Michael. :)
Quote from: striker on May 08, 2005, 05:25:28 AM
I do it once, and currently I do this....
mov eax, sizeof STRUCT
mov ecx, count_of_items_i_want
mul ecx
; go on and allocate 'eax' bytes
Do this appear correct ?
rgs, striker
You can save yourself one line of code with:
mov eax, sizeof STRUCT
mul count_of_items_i_wantAND, being done only once in your app, you would be wasting a zillion times more time trying to optimize those two lines compared to the fractional nanosecond you could save. ::) ::) ::) ::)
Raymond
Thanks for the replies.
I just wanted to know how to do it, not the fastest way.
Raymond - thanks for saving me 1 line. Ill use your approach.